- International Journal of Engineering Science and Application
- Vol: 5 Issue: 1
- Estimation of Optimal Range for Shape From Focus in Microscopic System
Estimation of Optimal Range for Shape From Focus in Microscopic System
Authors : Hülya Doğan, Elif Baykal Kablan, Murat Ekinci, Mustafa Emre Ercin, Şafak Ersöz
Pages : 27-37
View : 16 | Download : 11
Publication Date : 2021-03-31
Article Type : Research
Abstract :Shape From Focus (SFF) is one of the most preferred approaches in microscopic systems to reconstruct the 3D shape of the specimen. Classical approach generates a 3D shape using the 2D image sequence with the same field of perspective and different focused regions. In order to scan the specimen in traditional SFF approaches, the microscope platform is moved along the specified range, which is randomly defined between the begin and end locations on the Z axis. A certain amount of 2D image sequence with the same field of perspective and different focused regions are acquired in limited movements between these locations. However, the range and amount of 2D image sequence are effective to extract the entire 3D structure of the specimen, they must be optimized for the type of the examined specimen and magnification objective. In this study, a novel SFF approach is improved to scan the whole 3D structure of the specimen by estimating the optimal range between the begin and end locations of microscope platform. Unlike previous approaches, problems such as random amount of 2D image sequence with the same field of perspective and different focused regions, random determination of the range between the begin and end locations and the occurrence of outliers and noise around the 3D shape are solved in the proposed approach. Our experiments are performed on data sets collected from the light microscopy using specimen prepared for cytopathological examination. Qualitative and quantitative results demonstrate that better performance is achieved by our suggested SFF approach.Keywords : Focus measurement operator, Shape from focus, 3D shape reconstruction